Change management for AI adoption: agents only work once the change is led
How to lead the introduction of AI agents in sales along Wilfried Krüger’s phase model – from initiation to consolidation – and who is accountable for what in each phase.
The agent is live. It analyses customer data, prepares briefings and suggests next steps. Technically, everything works. And yet the sales team carries on as before: briefings go unread, suggestions are ignored, approvals pile up because nobody knows who is supposed to give them.
Situations like this show what introducing AI agents is really about. Technology is one part of it. The other part is the change in your organisation – and that calls for change management.
Why AI adoption rarely fails because of technology
When AI agents work in sales, tasks, roles and decision paths shift. Who prepares, who approves, who is accountable for the result? If these questions stay open, a working agent does not become an effective one.
Market observers see a risk here, too. In a forecast from June 2025, Gartner expects more than 40% of agentic AI projects to be cancelled by the end of 2027 – due to costs, unclear business value or inadequate risk controls. None of these reasons can be solved with better technology alone. All three demand clarity: about the goal, about the value and about accountability.
That is why companies often fail not because of technology, but because of a lack of clarity. This is exactly where change management for AI adoption comes in.
Krüger’s phase model
We draw on the phase model of change by Wilfried Krüger, described in “Excellence in Change”. It divides change into five phases:
- Initiation: the need for change is clarified, and the people who are to carry it are won over.
- Conception: goals are set and the measures that lead there are planned.
- Mobilisation: those affected are informed and involved; willingness and ability to change are built.
- Implementation: the planned measures are put into practice.
- Consolidation: the change is anchored so that it lasts.
The model fits the introduction of AI agents because it does not stop at go-live. It starts with the question of whether anything needs to change at all – and only ends once the new has become everyday practice. At REVAGO, the five phases map to the building blocks Evaluate & Enable, Go and Operations.
Initiation and conception: in the AI coaching
You lead the change, as leadership. Goals and mandate come from you – REVAGO does not set them. In AI-first coaching, you find out whether and how AI agents can secure and grow your revenue. That is initiation: you clarify the need for change before anything is built.
If the answer is yes, conception follows. At the end of the coaching you have a clear goal, a map of the way there and a mandate that defines the task and accountability for implementation. If the answer is no, that is a result too – and a more honest one than a project that gets cancelled later.
Mobilisation and implementation: clarifying roles in the human-in-the-loop
In mobilisation, the change becomes concrete for the people who work with the agents. With your process owners we design the processes – starting from the desired outcome, for agents and people who work together.
At the heart of it: for every task it is defined whether an agent
- prepares, for example analysing customer data and creating briefings,
- recommends, for example suggesting next steps and follow-ups,
- executes with approval, for example a reply in an ongoing conversation that a person approves,
- or acts autonomously within limits, for example with standard replies on defined paths.
This definition answers exactly the questions from the opening: who reads the briefing, who approves, who decides? That is how effective autonomy and efficient control belong together in the human-in-the-loop model.
Then comes implementation: we build your agents, connect them to the agreed systems and take everything live. How a clear mandate turns into a running process is the subject of the article on Go.
Consolidation: anchoring the change in operation
Many introductions end at go-live. For the change, that is too early. New ways of working take time to become second nature, and agents need to evolve with markets, data and processes.
That is why REVAGO runs the system on an ongoing basis: technical operation, maintenance and improvement within the agreed scope, aligned to your revenue goals. That is consolidation in the sense of the model – and the reason why an agent is never finished. Because we track the impact continuously, it stays visible whether the change is taking hold.
Your next step
Change management for AI adoption does not start with technology, but with a question: whether and how can AI agents secure and grow your revenue? For an overview of all five phases and how they fit together at REVAGO, see our page on change management for AI adoption.
If you would like to know where you stand, start with the short assessment: five short questions, followed by a conversation with a REVAGO coach.